It's been a while since I've sat down and shown the process of making a puzzle, mainly because I've been busy in the shop working on making puzzles, and haven't had time to write, however the Really Bent Board Burr by Derek Bosch is one that is worth writing about.
I've never owned a copy of this puzzle, and it has always intrigued me. I've talked about making copies for long enough, and now I finally have, so here's a little bit of insight to the puzzle, and the process of creating it. The puzzle was originally produced by Tom Lensch back in 2007 and the craftsmanship as you'd expect was superb. Hopefully I'm able to do it justice, but I'll let you be the judge of that.
Before I get into the details of making this puzzle, here's a few interesting things about the puzzle. It's hard to tell from the assembled puzzle, but this is a 6 piece puzzle, with three different types of pieces used in the construction. Each piece forms a 'Z' shape with two 'C' pieces attached to a central 'O' piece. The puzzle itself has two different solutions using the same 6 pieces. An easy and a hard solution. The easy solution is a level 10.6.1.4 while the hard is a staggering 20.2.10. In all honesty calling the easy solution easy is a joke. This is a really tough puzzle both to assemble and dis-assemble, however the final shape is well worth the time to solve. (And no, I'm not smart enough to assemble it without help!)
The pieces form a set where three of the Z's have the C's attached with the opening facing in opposite directions, two with the opening in the same direction, and the final is a mirror image of the second piece type. Given the length of the C's and the minimal gluing surface to the central structure, the joints need to be reinforced to prevent them from breaking. Equally, the end of the C's need to be reinforced to prevent them breaking too. All told there's a lot of work to producing such a puzzle, however the end result is in my opinion worth the extra effort.
The start of any such puzzle is with the preparation of the stock. Square sticks need to be accurately milled from the boards giving straight, consistent sticks as a starting point. For this run of puzzles, I had a selection of Maple, Walnut and Lacewood to work with. Fortunately all the stock I had was 8/4, meaning that I could create sticks that were over 1/2" in diameter, resulting in a very pleasing and not small final puzzle. Overall, each puzzle requires 12 feet of wood to make, and 2 feet of dowel to pin the pieces ensuring they are strong enough. That's a lot of wood!
You can see my cheat sheet in the image above, where I mapped out the pieces and produced a cut list for the individual sticks needed to create the final puzzle. The colouring on the pieces on my cheat sheet is partly to make seeing each piece easier, but it also helps with the wood selected for each piece, resulting in a pretty interesting final puzzle piece.
The size of each of the pieces is determined from the width of the square stock. That sets the size of a single cube, and from there the units required are 1x1, 1x2, 1x5 and 1x7. I made a set of these pieces, which are easily created my combining the smaller units, all cut from the stock I'm using to ensure their size is accurate. You can see them in one of the photos below, sitting on my saw.
The square sticks are cut to the correct lengths for each puzzle in batches, and then stacked to create the pieces for each puzzle. In total, I cut enough pieces for 10 copies of the puzzle to be made. As I've mentioned before, once the jigs are setup to make the cuts required, the effort to create 10 copies is not significantly more than to create one, so it just makes sense to make multiple copies. I'm sure there are people out there who will be interested in a copy.
From each pile of sticks, the individual components of each individual puzzle piece are created. Since each piece consists of two C's and an O, those can be created en-mass, and then assembled into the correct puzzle piece.
After the individual components are completed, they are glued together to form the final puzzle pieces, and then holes are drilled through the O's to allow dowels to be glued into place, forming a much stronger joint between the components and helping to ensure that significant force would be needed to break the pieces. At the same time, the pieces are run across the table saw to add a flat bottomed groove in the ends of the pieces to allow a spline to be added. That spline will reinforce the ends of the C's again helping to ensure that the pieces will not break when the solver is playing with the puzzle. These ends are very weak without some additional support since there is very little gluing surface, but lots of force available given the length of the pieces.
Once the glue has dried, the dowels are trimmed with a special saw which does not mark the surface the blade rests against, and the ends of the c-pieces are rounded to both clean up the spline, and add a visual element to the puzzle pieces in the assembled state. The other advantage to doing this is that is hides any tearout that was created from cutting the groove in the ends of the pieces. Normally I will back-up the cut with another piece of wood against the back of the piece where the blade will exit. This prevents the wood fibers which are unsupported otherwise from being ripped away from the piece, however with this type of cut that is very difficult, and taping the joint is only partially successful. So from my perspective as a craftsman, this rounding is both useful and pleasing to the final puzzle piece.
At this point, the puzzle can be tested to ensure that all the pieces fit together. Unlike many other puzzles I've made there's no way to test the puzzle sooner. That means that I could have spent 10 hours making the pieces, and have nothing but scrap to show for it. Unfortunately, the pieces are so long that without all the dowels and splines, they are not strong enough to be put together into the final puzzle meaning that it's an all or nothing build. Fortunately with the experience I have gained over the last few years, the puzzles went together without issue. Only minor sanding was required in a couple of places on one of the puzzles to ensure a good fit.
With the pieces tested and fitting together, they can be final sanded up to 600 grit to ensure a smooth and tactile surface, then finish can be applied to bring out the beauty of the wood. My go-to finish for puzzles is still a thinned lacquer then the Beall Triple Buff system to really make the pieces shine.
That's about all there is to it. Each puzzle takes around 15 hours to make from start to finish, and I'm now very happy to have one of these excellent puzzles in my collection. They are a lot of work to make, and I'll be honest, as happy as I am to have one in my collection now, I'll not be making more of these any time soon! Hopefully the write-up was interesting, and hopefully I'll be back to writing more soon.
I've reviewed a couple of Michael Toulouzas' puzzles and each time I receive a new puzzle from him, I'm stunned by the look and quality of his work. The latest piece I received is "The Illusion Puzzle", a six piece interlocking puzzle, that as Mike himself points out is not quite what it seems.
By the time I saw this puzzle, and decided that I wanted a copy, Mike had already sold through his initial run, but agreed to let me know when he had more available. It took a little time, but true to his word, he got in touch to offer me a copy. I didn't hesitate, and before too long, the box arrived from Mike.
My initial exploration didn't reveal much, and the puzzle seemed to be fairly well locked up. Of course with some careful finger placement, I found a little movement, and before I knew what had happened, I had six dissimilar pieces in my lap, and no idea where they went in relation to each other, and no real idea about the motion needed to put it back together. Guess I was going to have to truly solve this one with no hints, or reference from the solved puzzle.
Looking at the pieces, there's some interesting pyramids which stick out and really do a great job of getting the way of solving the puzzle. That said, they also give you a clue as to how each piece must be oriented in the solution, since those blocks fit into cutouts in the other pieces. With some analysis, it's possible to minimise the possible combinations and significantly reduce the permutations you need to try. Of course there's only one way that the pieces will come together, and I had many attempts where I thought I had the orientation correct to be thwarted by one of those pyramids stopping the pieces from coming together.
Despite the puzzle exploding when I took it apart, this works very similar to a Sliding Star puzzle, and two halves glide together smoothly once the pieces are in the correct orientation, and you have the correct two sub assemblies created. I can now take it apart and put it back together without it exploding, and each time marvel at the design which created such a complicated set of interactions in what looks like a simple puzzle.
If you like this type of interlocking polyhedral puzzle, I highly recommend The Illusion, or any of Mike's work, if you can get a copy.
Cube-16 is another Stewart Coffin design which is an improvement on his earlier design Patio Block (STC#82). Cube-16 is numbered STC#205 in Stewart's numbering scheme and the goal is to take apart the cube, and then return it to its original state. My copy was made by John Devost, and I was lucky enough to pick up a copy when he recently offered them on Puzzle Paradise.
The external appearance of the cube is identical to the earlier Patio Block design so it would be easy to confuse the two until you pick the puzzle up. Patio Block was an eight piece non-interlocking puzzle, so without a box to hold the pieces, it would come apart easily when you tried to move it. Cube-16 on the other hand is a fully interlocking cube, so you can pick it up with no problems. Of course with it being interlocking, finding the first piece to remove can be a challenge.
I'm not aware of too many copies of this puzzle being out there, so when John Devost announced that he's made a few copies I jumped at the chance. John hadn't been making wooden puzzles for a long while, and many of us in the puzzle world thought he'd hung up his tools and given up. I'm pleased to say that it looks as though he's back, and making some great puzzles again. Welcome back John.
The puzzle itself measures just under 2" x 2" x 2", and is made from Afzelia Burl. As you can see, the appearance of this wood is stunning with beautiful waves and swirls and eyes throughout the pieces, but what you can't see is how the wood smells. This is an amazingly fragrant wood, even after it's been lacquered and polished by John. I really can't describe quite how strong it is. The best I can do is to say it's like walking into a loose leaf tea shop and smelling that sweet aroma. Yes, it's that strong. The fit on my copy is good. John mentioned that it was a little tight when he had finished making it so he shipped it with a couple of spacers marking the key piece to make sure that I didn't end up breaking anything trying to find the first move. Hopefully after a little while here in California, it will loosen up nicely and there should be no problems with the fit.
While the Patio Block design was an eight piece puzzle, Cube-16 is a five piece puzzle, but that doesn't mean to say it's significantly easier than the original design. I should probably clarify 'original' here. Even Stewart's original Patio Block was inspired by an even earlier puzzle. The inspiration for the Patio Block design was a ten piece puzzle created by Toshiaki Betsumiya, and another similar puzzle which was an eight piece version by Kevin Holmes. All of these designs came out of studies to create 4x4x4 cubes with external symmetry. Stewart took those ideas and created the Patio Block, and later Cube-16.
Each of the five pieces are unique as is the case with many of Stewart's puzzles which adds to the challenge. That said, I wouldn't say it's that hard. I expect that most people will be able to find the solution in around half an hour making it a very approachable puzzle.
I'm glad I was able to get my hands on a copy of this puzzle, and John has suggested that it would be a good design for me to make a few copies of. I may just have to do that, although I doubt I'll have any wood which will look quite as stunning as the Afzelia Burl.
Wow, doesn't time fly. It's nearly the end of 2012, and I realise that I still haven't sat down and written up the Tetraxis puzzles that I received from Jane Kostick over a year ago. Worse that that, I created the video review back in October (last year)! Not only that but after being at IPP and playing with more of Jane Kostick's excellent puzzles, I've bought more from her and really need to review that too. So with all the procrastination out of the way, I hope you enjoy the review of this fine puzzle set.
Some time before IPP31 I came across Jane and John Kosticks website, and really liked the look of their puzzles. Digging around a little, I found that as well as the mass produced plastic versions of the puzzles, Jane also made some rather unique versions from wood as well. So chancing my arm, I got in touch and asked if she would make me a set.
Jane and I chatted back and forth for a while via email, and eventually agreed on some woods to use from Jane's fairly large stock, and she set about making the puzzle pieces for me. I have to admit that I really enjoyed talking with her about the puzzle itself, how the magnets were attached to the sticks, her experiences with different woods, and many other topics, including IPP. In the year that has passed since then I still talk to Jane via email occasionally and on returning from IPP 32 I threw her another email commenting both on the great puzzles that she had in the Design competition, and also a particular puzzle that John Rausch had given me to play with, but more on that later.
Unbeknown to me, fellow puzzler Allard was also having a set of Jane's sticks made and unlike me he wrote about them in a much more timely fashion.
My copy is a three layer puzzle with a rhombic triacontahedron in the centre, then Jane and John's Six-axis in the middle and a Ten-axis Tetraxis frame on the outside. The inner parts are made from Bubinga, Black Palm and Maple, and the outer Ten-axis is Cherry. They have trademarked the name Tetraxis, and use it to cover all of the puzzles they make including the Tetraxis Star, Tetraxis toy/puzzle, and Tetraxis magnetic sculptures in wood.
The whole structure shows the fantastic geometric relationships that are found in John's stars and really helps you to visualise the geometry in play. So many puzzles are based on the geometry here, and you'll probably recognise the shape of the six-axis Tetraxis as the same shape as Stewart Coffin's Jupiter puzzle (amongst others).
When I originally ordered the puzzle from Jane, things took a little longer than she would have liked due to some dull saw blades, so she ended up sending me one of her husbands Tetraxis stars (4 axis) as a little extra to say sorry for the delay. Now given that this was a custom order I really don't think it took long at all, and the craftsmanship is superb. To my mind there was no delay and really it wasn't an issue at all, so thanks for the little extra Jane, it's beautiful too.
As you can see from the closeup pictures, the fit and finish of the sticks is excellent, and Jane has even signed the pieces which on this scale is no easy task. The overall structure may be fairly large as you can see in the video, measuring 4" however each stick is just 0.25" thick on the outer Ten-axis assembly.
Putting the pieces together isn't overly challenging, and the magnets pull everything into place nicely so in terms of a challenge, this isn't the most difficult puzzle you'll play with, however it is remarkably relaxing to just start putting the pieces together and hear them snap into place. The resulting shape has a sculpture like quality which sits very proudly on my puzzle shelves, and gets a lot of attention from visitors.
It's a beautiful puzzle, and truly a work of art. This year at IPP, Jane and John had a number of new designs entered which I was fortunate enough to play with. Not only that but John Rausch had a one off puzzle Jane had made which he handed me to play with. Having talked to Jane about this puzzle after IPP, I now have a copy of that too, so come back tomorrow to read about her IPP entries and a new design that I fell in love with.
After the quick interlude as I prepare for IPP, here's the third part of the Hex Stair saga. Despite only having 11 puzzles made in the last month, it feels like I've been working on this project a fair bit longer. Still seeing all the puzzles finished and ready to go to IPP with me I'm really pleased with the results.
I'm not sure how anyone else views the finishing process, but for me it starts long before you ever get out a brush or some lacquer. Much of the look of a finished puzzle, or any wooden object really comes from the choices you make when you're putting it together. There are subtle details which really help to 'finish' a project. For the Hex Stair (and the Domino Tower) puzzles, adding the very slight 45 degree bevel to the edges of the pieces really adds to the overall look. Without it, the puzzle looks incomplete. So for me that's the first part in the finishing process. After each of the blocks are cut to size, I add a tiny bevel to each piece. It's a time consuming process, but without it the pieces just lack that little edge that they'd otherwise have.
Getting ready, each of the puzzles are assembled, as I will be sanding only the outside of the pieces. The reason for this is that I send a lot of time ensuring that the pieces are all the exact same size, to ensure a tight fit on the assembled puzzle. If I were to sand the pieces, then I'd lose that fit, and the puzzle would become too loose, or not fit at all! You'll remember that I aim for one thousandth of an inch tolerance between pieces. Sanding will remove much more than that!
Given the finish I get from the blade of the saw, you could ask why sand the puzzle at all? The main reason is the feel of the puzzle in your hands when it's sanded. Although the puzzle if perfectly dimensioned, the feel of the wood can still be a little rough. By working up through the various grit of sandpaper, we'll take the wood to being silky smooth to the touch, and make it something that you'll want to hold. Given that I'm starting off with something which is close to a finished surface, I'll start sanding at 220 grit, then move up to 320, 400, and eventually 600 grit. The last two grits are more polishing the wood than removing imperfections, so very light passes are all that's required at that stage.
With each of the sides, and both top and bottom sanded, the puzzle is left coated in a fine sawdust. You'll notice that I'm working with the sandpaper attached to my granite block. I work the puzzle across the paper rather than take the paper in my hand and bring it to the puzzle. The former ensure that the surface is dead flat, and I don't over sand any particular area, where there latter, the different pressure from my fingers would lead to imperfections. Before moving to the next grit, the dust has to be taken off, otherwise it will be ground into the surface of the wood, and you will end up working harder to get the surface you're looking for. To do this I use two processes. First up, I have an air compressor with a fine nozzle on it. Using that at about 60 PSI, I blow most of the dust off the surface, taking care to ensure I get any dust out of the pores of the wood. On wood like the Paduak I'm using which has fairly deep pores, the air easily clears them out.
Once I've blown the dust and cobwebs away, I use a Tacky cloth to take care of anything that's left on the surface. The Tacky cloth is has an almost waxy feel to it, and does a great job of taking anything loose off the surface. With that done, I can continue up through the grit until I finally reach 600. All in all it takes about forty minutes per puzzle, but since I was working with all the puzzles, around 3 hours in total.
It may be a little hard to tell the difference from the photograph, but this is the puzzle sanded up to 600 grit. The real difference is in the feel of the wood. Now much smoother, the finish is almost like glass.
Next up in the process is to apply the finish. I use a three stage process currently. First up is to apply a couple of thin coats of lacquer. I mix the lacquer 50/50 with thinner, and apply two coats to the puzzle pieces. Working with the lacquer thinned like this, I have found I don't end up with runs or drips. Given the size of the pieces I'm working with I use a small brush to apply the finish which could leave brush stokes with a thicker mixture. Each coat is applied and allowed to dry overnight before adding the next coat in the morning.
It takes around 20 minutes per puzzle to apply a coat of finish, and then it's left to dry. I'll show side by side photos below, so you can see the difference after each stage. It will probably not be too obvious, however the before and after shot shows worlds of difference!
After the two coats of lacquer are applied, I take a good look at each of the pieces. Sometimes the wood absorbs the lacquer more in certain areas, and the finish can appear uneven. If that's the case, I'll go back and apply a third, or even fourth coat of lacquer until the wood has absorbed the lacquer evenly. After each coat, the lacquer is left to soak in for around thirty minutes, and then I'll come back and rub off any excess with a clean cloth. If the lacquer pools on the surface, it will dry hard and uneven, which can affect the fit of the puzzle, and certainly doesn't make it look any better!
Once I'm happy, I'll apply a liberal coat of the Watco finishing wax. This helps the pieces to slide past each other, and adds another layer of protection for the puzzle. After all these will be played with, so I want the wood to be protected. I leave the wax on the puzzle for around 15-20 minutes, then with another clean cloth, rub the excess off. Part of this process I also buff the pieces as I work the wax into the surface, but mostly I'm removing the excess.
The final part to the finishing process is to apply a coat of Renaissance Wax. This incredible substance (which is not cheap!) brings up an amazing shine on the wood. Applying it leaves the wood with a slightly tacky feel, and a finish which is less than mirror. I apply the wax as evenly as possible, and then let it sit for 20 minutes. After that I take a clean cloth and start buffing the surface. It takes about 20 minutes per puzzle, but the wax really polishes up the surface and starts to make the wood shine.
After the initial buffing, I take the puzzle apart, as the wax gets pushed into every little gap. This needs to be cleaned out before the puzzle is re-assembled and given another buffing. All told the process takes nearly 45 minutes per puzzle, but as you will see below, the results are worth every minute of it!
As you can see the difference from start to finish is dramatic. The surface ends up being quite reflective, and really brings out the grain in the wood. It may take ~4 hours per puzzle but the results speak for themselves.
The images below show each stage of the process compared with the unfinished puzzle. It may not be too easy to see the difference, as the changes are subtle.
I hope you've enjoyed the writeup of my finishing process. I freely admit that I'm no expert, and I'm learning as I go, however several of my readers have asked so hopefully this is useful to you. This is by no means the definitive guide to finishing, and certainly isn't appropriate for all applications, however it does work for me when finishing puzzles, and I'm happy with the results. From the feedback I've had from those who've bought my puzzles, they seem to agree that I'm doing something right!
Following on from by first post on the Hex Stair (Part One), I'm moving from the initial copy which I made for myself and realised it was far too big, onto a smaller version, which I'll be making a small run of puzzles to sell. Making things smaller adds new challenges so read on to see what I had to do to overcome them.
As you can see above, I have seriously scaled down the size of the pieces, which makes for a much more manageable puzzle size. I am slightly torn in all honesty; I love the big solid chunky copy I made initially, but also appreciate that it's far too big for most people, and the more compact size is far easier to work with ... or is it?
Having decided on the dimensions for this smaller version, I cut and milled my stock, cutting what I hoped would be enough sticks to make a reasonable run of puzzles. The pile of sticks looks like a lot, but I have no doubt that I'll get through them pretty quickly. If you're interested, the woods are (from left to right) Paduak, Wenge, Curly Maple, Purpleheart (on top), Birdseye Maple, Red Palm, Paduak.
With everything setup, I started batching through the cutting of the pieces, and despite needing 42 pieces per puzzle, once everything is setup, this goes fairly quickly. I keep my digital calipers beside me and keep checking the cuts as I go to make sure I've not had any errors introduced, as the biggest reason I have found for a puzzle not fitting correctly is tiny differences in the tolerances of cuts. Anything more than about five thousands of an inch between pieces and the fit will not be good enough. Ok, five thousands is pretty small I hear you cry, but I try to get my pieces to less than two thousandths to make sure I don't have problems. Sadly I've learned from experience that even small margins like this make the world of difference and cause a lot of frustration when gluing pieces together in the type of puzzles I'm making.
Having batched out a good few pieces; enough to make a few puzzles; I take a break from cutting the pieces and move to the router to add the bevel to the edges of the pieces. I find that taking a break like this keeps me focused and alert, rather than becoming complacent as the motions get repetitive, and it's all too easy to lose focus ... and as I have already experienced, a tiny lapse can be very costly!
With all the pieces cut and beveled, it's time to start gluing the puzzles together. You'll remember the crude jig that I made for the initial puzzle, which worked pretty well. I found out though that with the smaller pieces, there's not a lot of gluing surface, and I made the pieces almost as tall as the are wide. (They're not perfectly square!) Because of this, it's easy for the pieces to get misaligned, so I felt I needed a better jig...
As you can see this jig is a little more advanced than the original, however the main drawback is that it will only work for this puzzle, and only for pieces cut to the exact sizes that I have used. While that may seem like something of a waste, for the most part, the jig is made from offcuts from the sticks I used so really it's putting small cuts which would otherwise be used in my fireplace to good use. The jig is a very snug fit for each piece, and as you can see each piece is well supported meaning that each completed piece I make in the jig will be identical and it's also very quick to use, since there's no way to misalign a piece. With this jig, it takes around 15 minutes to make each individual puzzle piece, meaning that I can make a complete puzzle in around 2 hours (allowing for the curing time of the glue). This as nearly 2.5 times faster than previously. While it may seem like it still takes a long time, I'd rather take my time than rush and end up with a useless pile of firewood. After all, a high quality puzzle doesn't get made in a minute.
With the jig doing all the hard work for me, it doesn't take too long to make a copy of the puzzle, to the point where the outside faces get sanded and then finish applied. In Part Three, I'll cover some of the finishing process.
One thing I noticed when assembling this version of the puzzle is that it's actually easier to put together than my original version. One of the reasons is that I've found a particular rotational move which allows the alignment of the pieces to happen much more easily. I didn't find this on my original version, I think partly because it is much more squat than the new dimensions. The extra height makes it easier to do this move (although I have gone back and found that it is also possible on the original copy).